Data center automation is about cutting losses, not layoffs
Data center automation has a reputation of being in the "if you have to ask, you can't afford it" category. It also carries the assumption that it will eventually make in-house IT obsolete--but both assumptions are wrong, writes IT manager Jasmine McTigue from medical equipment maker Carwild Corp.
"It's easy to see why some of us worry. In manufacturing, automation translates directly into staffing cuts--if Detroit auto assembly line workers could have nixed the robots, they sure would have," McTigue writes. But the real advantage of IT automation is in detecting when something has gone wrong and responding quickly. "A 'faster than human' action may in itself generate a concrete ROI, depending on the transactional load and value of the application in question," McTigue adds.
That's where data center automation begins to look more affordable. By calculating how much business is lost per minute when particular servers go down--say, the ones that process customer orders--you can produce a metric that puts an actual value on lost time responding to problems. (Some customers will come back if they can't order because a server is down, but even calculating the loss conservatively gives useful dollar amounts.)
Applying that basic cost-per-minute statistic to actual historical outages, and calculating how much wouldn't have been lost with an inhumanly fast response to the problem, makes business sense of the proposed automation. ("Last year we had seven midnight outages that cost us an average of $5,000 each in labor. By automating this system, we could avert 75 percent of those outages, saving $26,250 per year. The automation necessary would cost $15,000.")
Those hard numbers are exactly the kind of concrete value proposition that a CEO and CFO want to hear.
- see McTigue's post at InformationWeek